A review of some of Jeff's work at the Recurse Center, including a couple of image compositing programs, plus a demonstration of the Moldable Development approach to explore the Docassemble UI scripting framework for helping people file legal papers online by collecting necessary information in a structured way (e.g. for requesting a protective restraining order).
Marc's invited friend Gene Bellinger demonstrating a tool for working with dynamical systems diagrams, using large language models (ChatGPT-4) as a way to fill in nearby context for the model diagrams.
Thompson Morrison's use of large language models in a particular way, "describe how these things are connected", where he has some understanding of the connection which he's studied, and the LLM text may produce some additional angles that could become draft pages in his knowledge graph.
Thompson Morrison's exploration of how to better serve "noncompliant learners" in the classroom who have more sophisticated needs for how the material is presented and worked with.
Systems dynamics, dynamical systems diagrams, "archetypes" in dynamical systems with names for productive or destructive patterns composed of coupled reinforcing and stabilizing loops.
AI (LLMs) and what they systematically fail to capture, which is closely observed insights (versus common patterns) -- Eric's observations about automatic textual summaries of communications related to software systems incident resolution -- the summary is good at describing events, and can completely miss the key turning point in resolution.
Learning from Incidents, complex software ecosystems, programmer mindset ("inventing the future") may act against getting grounded in the circumstances which produced the incident.
Residuality Theory (the architecture of a system is the residue of its responses to survive the forces which have challenged it)
10:51:33 From Eric Dobbs Barry O’Reilly, Residuality Theory & my notes from a conference talk on the subject http://wander.dbbs.co/residuality-theory.html
History of agile, C3, special talents of various C3 members, Fred Brooks on No Silver Bullet, the Mythical Man-Month, learnings from the IBM OS/360 project which informed later work.
Gerald Weinberg, The Psychology of Computer Programming; anecdote (which I couldn't find right away) about a debugging process, where the key insight comes from someone who is listening to the other team members converse, and then asks a specific question -- less visible as "busy" but closely inspects the issue and provides the key insight.
The importance of getting to know your collaborators at a conference, in your field, or in a software team; challenges about "is this communication unproductive" versus "it will pay off later to know who I need to talk with about things".
Difficulties of making these connections on a remote team, unless the organization supports it.
Games as a way to both develop relationships among participants, and also illustrate emergent behavior (Ward's descriptions of breaks for social games hosted by an expert, and their value to the PLoP conferences).